USER FORUM
(you are viewing a thread; or go back to list of threads)
Modeling Gridfinity (by S Groesz)
How best can the gridfinity profiles be modelled?
Design specs: https://gridfinity.xyz/specification/
The model itself is pretty simple. The complication related to Solvespace is the rounded corners with 45 degree bevels.
For the purpose of this exercise, I am focusing only on the baseplate design.
The method I can think of is to sketch the profile of the 4 sides (of the baseplate) between the bevels, extrude each side, then us the revolve tool to extrude the corners. But the method I have in mind would require many layers.
- One initial layer to define the 42x42 outline with the rounded edge
- One sketch layer per each side (+4 layers) plus an extrude for each side (+4 layers)
- One sketch for each corner to be revolved (+4 sketch +4 revolve)
So, 17 layers (groups)? Is this the best way to achieve this?
Another method I can think of would be to extrude the rough 42x42x5 shape then cut the solid to the correct specifications. However, I'm not certain this would be any easier or better to simplify the model.
Attached is what I have so far which gives the rough dimensions of the shape. I was sort of hoping I could duplitranslate the solid models, which would reduce the number of sketches. That is, sketch the first side, extrude, step-rotate the resulting solid to create the 4 sides, then do the same for the rounded corners. But it doesn't seem I can duplicate an extrusion (?).
Design specs: https://gridfinity.xyz/specification/
The model itself is pretty simple. The complication related to Solvespace is the rounded corners with 45 degree bevels.
For the purpose of this exercise, I am focusing only on the baseplate design.
The method I can think of is to sketch the profile of the 4 sides (of the baseplate) between the bevels, extrude each side, then us the revolve tool to extrude the corners. But the method I have in mind would require many layers.
- One initial layer to define the 42x42 outline with the rounded edge
- One sketch layer per each side (+4 layers) plus an extrude for each side (+4 layers)
- One sketch for each corner to be revolved (+4 sketch +4 revolve)
So, 17 layers (groups)? Is this the best way to achieve this?
Another method I can think of would be to extrude the rough 42x42x5 shape then cut the solid to the correct specifications. However, I'm not certain this would be any easier or better to simplify the model.
Attached is what I have so far which gives the rough dimensions of the shape. I was sort of hoping I could duplitranslate the solid models, which would reduce the number of sketches. That is, sketch the first side, extrude, step-rotate the resulting solid to create the 4 sides, then do the same for the rounded corners. But it doesn't seem I can duplicate an extrusion (?).
(no subject) (by Paul)
This might give a hint. One group was having NURBS problems, but I went ahead with the step rotating anyway and it seems to have cleaned up.
(no subject) (by Paul)
It might help if I attached the file...
Looks great! (by S Groesz)
Thanks for this, Paul! It looks great. I'm going to study it and hopefully everything will be clear. But this is exactly what I was looking for.
I didn't even get as far as having nurbs issues, but I had already expected them and planned to get around that by reverting to a triangle mesh.
I didn't even get as far as having nurbs issues, but I had already expected them and planned to get around that by reverting to a triangle mesh.
(no subject) (by S Groesz)
I finally had a chance to review your file in depth. It appears that the trick is to generate the first extrude as two-sided (?).
When doing rotations, I feel like https://imgur.com/face-keyboard-smash-u3lWw until I get the result I want. By the time I'm finished, I don't have enough brain matter left to think about how I actually got it to work in the end.
Your file showed me that it is possible, so I created the file from scratch so I could learn. Unfortunately, the slvs files don't indicate which objects (point and line) the rotate step was based on. This leads me to "mash the keyboard" (try different point/line combinations) until it works. It came together more easily when I realized your extrude was two sided, which placed the origin point at the middle of the extrusion.
When doing rotations, I feel like https://imgur.com/face-keyboard-smash-u3lWw until I get the result I want. By the time I'm finished, I don't have enough brain matter left to think about how I actually got it to work in the end.
Your file showed me that it is possible, so I created the file from scratch so I could learn. Unfortunately, the slvs files don't indicate which objects (point and line) the rotate step was based on. This leads me to "mash the keyboard" (try different point/line combinations) until it works. It came together more easily when I realized your extrude was two sided, which placed the origin point at the middle of the extrusion.
(no subject) (by S Groesz)
My final result
(no subject) (by S Groesz)
Here's my attempt at modeling the base of one of the bins. I followed the same work flow. I disabled nurbs at the first rotate step. I had to duplicate the extrusion 4 times in the rotate step so that I could get the result I wanted. Ideally I would just need to duplicate the model once and rotate the copy 90 degrees, which would give me the profile of all 4 sides. Having the 4 models overlapping might be causing more complication than necessary?
(no subject) (by S Groesz)
Here's an alternative method I came up with which doesn't require overlapping solids. I start with the major profile, then fill in the missing sides and then the rounded corners.
It seems counter intuitive to me, but it makes sense, that this file is much larger than the first attempt.
This second attempt I was able to leave nurbs enabled.
Is one better than the other? With regard to software performance is what I'm mostly concerned with. These are the most basic profiles used for Gridfinity. The complicated geometry occurs when we start adding cutouts for whatever widgets need to be stored...
It seems counter intuitive to me, but it makes sense, that this file is much larger than the first attempt.
This second attempt I was able to leave nurbs enabled.
Is one better than the other? With regard to software performance is what I'm mostly concerned with. These are the most basic profiles used for Gridfinity. The complicated geometry occurs when we start adding cutouts for whatever widgets need to be stored...
(no subject) (by S Groesz)
When I combined the base and the bin base files, it was apparent there was an error with the bin base model. The drawing has an error. I blindly copied the 35.6 dimension shown on the drawing for the dimension of the bottom surface. The correct value is 34 or maybe 34 - 0.5 = 33.50.
This is why I like solvespace - it's relatively simple to model something to find the missing dimensions. It also serves to validate the source drawing...
Luckily with the alternative file I am able to change this value very easily. The original model I made seems to fall apart when I try to change this value and then repeat the transuplicate step.
This is why I like solvespace - it's relatively simple to model something to find the missing dimensions. It also serves to validate the source drawing...
Luckily with the alternative file I am able to change this value very easily. The original model I made seems to fall apart when I try to change this value and then repeat the transuplicate step.
Post a reply to this comment: